After euthanasia, pancreas were removed and fixed in phosphate-buffered formalin 10% (phosphate buffer pH = 7·2) for 24 h. The organs were conserved in alcohol 70% until histological processing and paraffin inclusion. Five-μm sections were cut and stained with haematoxylin and eosin (H&E). All islets on the slides were analysed and the following criteria

were employed to determine insulitis score: 0 = intact islet; 1 = peri-insulitis; 2 = moderate insulitis (< 50% mononuclear infiltration); and 3 = severe insulitis (more than 50% mononuclear infiltration). Spleen cells were cultured in RPMI-1640 medium supplemented AZD4547 molecular weight with 10% fetal bovine serum, 2 mM L-glutamine and 40 mg/l of gentamicin and then plated at 5 × 106 cells/ml in 48-well flat-bottomed culture plates (Nunc, Sigma-Aldrich) and stimulated with 10 μg/ml of recombinant heat shock protein 65-kDa (rhsp65). Cytokine levels were evaluated 48 h later by enzyme-linked immunosorbent assay (ELISA) in culture supernatants using interferon (IFN)-γ, interleukin (IL)-5 and IL-10 BD OptEIA Sets (Becton Dickinson, San Jose, CA, USA) and tumour necrosis factor (TNF)-α

Duoset (R&D Systems, Minneapolis, Nutlin 3 MN, USA). The assays were performed according to the manufacturer’s instructions. Spleen cells were collected, the red blood cells were lysed with Hanks’s buffer containing NH4Cl and the remaining cells were adjusted to 2·5 × 106 cells/100 μl. These cells were incubated with 0·5 μg of fluorescein isothiocianate (FITC) anti-mouse CD4 (clone GK1·5) and 0·25 μg of allophycocyanin (APC) anti-mouse Suplatast tosilate CD25 (clone PC61·5) for 20 min at room temperature. Staining for FoxP3 was then performed utilizing the phycoerythrin (PE) anti-mouse/rat FoxP3 Staining Set (eBioscience, San Diego, CA,

USA), according to the manufacturer’s instructions. After incubation, the cells were fixed in paraformaldehyde 1%. The cells were analysed by flow cytometry using FACSCalibur (Becton Dickinson) and BD CellQuest Pro software (Becton Dickinson, San Jose, CA). Results are presented as mean ± standard error of the mean (s.e.m.). For diabetes incidence, the χ2 test was used. In all other cases, one-way analysis of variance (anova) was used for parameters with normal distribution and the Kruskal–Wallis test for parameters with non-normal distribution. Dunn’s test was used when necessary. Significance level was P < 0·05. Statistical analysis was accomplished with SigmaStat for Windows version 3·5 (Systat Software Inc., Chicago, IL, USA). Weight variation, glycaemia and the score of mononuclear infiltration in the pancreas were analysed in mice immunized with BCG alone or with prime-boost (BCG followed by pVAXhsp65) before diabetes induction with STZ. As shown in Fig. 1a, although all the groups gained weight, BCG–STZ and BCG/DNAhsp65–STZ exhibited a smaller variation (3 and 1%, respectively) in comparison to the control group (9%).